GET-Evidence

Take care to merge the rsid pages with the amino acid change pages so that they do not overwrite each other.

Cases:

1. AA page exists with rs# association, rs# page does not exist, robot knows rs# but not AA -- should automatically edit the AA page instead (done(*))
 2. Robot knows rs# but not AA, not AA/rs# association available -- robot should add/edit rs# page
 3. Robot imports a new variant with AA and rs# (genome importer is currently the only way AA/rs# association is made), rs# page exists, AA page does not exist -- robot should rename rs# page to AA
 4. Genome importing robot imports a new variant with AA and rs#, rs# and AA pages already exist (e.g. AA page created by user) -- try automatic merge, or warn editors that a manual merge is needed (incl. a warning on the pages themselves?)

The merge/rename stuff should be done by a separate process, rather than building that logic into the genome importer. (Perhaps tie in with the trait-o-matic data processing side -- e.g. dbsnp->aa change can be a quick db lookup instead of the full gff_nonsynonymous_filter process.)

(*) Need to fix handling of "one rsid -> multiple nsSNPs" case.